This paper presents a comparative study between Polynomial and V-transform coefficients for the fault classification in analog filter circuit using different kernel functions of Support Vector Machines (SVM). V-transform is a non-linear transformation which increases the sensitivity of polynomial coefficients with respect to circuit component’s variation by three to five times. It makes the original polynomial coefficients monotonic. Support Vector Machine is used for fault classification in polynomial and V-transform coefficients. The classification accuracy in both Polynomial and V-transform coefficients are increased by varying the kernel parameters c and epsilon associated with the use of SVM algorithm for the different kernel functions. The SVM’s are estimated in comparisons with the varied kernel functions by applying to the two feature sets. It is shown that the Pearson VII kernel function (PUK) provides good classification accuracy for the two feature sets compared to the other kernel functions such as Polynomial kernel function (POLY kernel) and the Radial Basis Function (RBF) kernel functions